Objections to the Nonrealist—and Replies

Scientific realism has been the subject of hundreds of papers and dozens of books. I have space to consider only a small selection of objections. The reader might imagine the following as the text of an extended interview.

a. Realism, antirealism, nonrealism

i. You reject scientific realism, so that makes you an antirealist. Why, then, do you insist on calling your position nonrealist?

Because I am a nonrealist, not an antirealist, although admittedly on the “left,” anti-realist side of nonrealism. A person who rejects theism is not automatically an atheist, for agnosticism is a respectable alternative. The same holds for scientific realism. Nonrealism is agnostic concerning the debate between realists and non­realists (Fine 1986; van Fraassen 2002). Thus I do not say that current theory and practice are false regarding the existence of entities and processes and truth claims about them. I do not deny that black holes and quarks and mirror neurons exist. But neither do I accept that current claims about them are the final, representational truth about these matters, or nearly so. I don’t see the point of doing either.

I do not object to “intentional realism,” whereby scientists (and science com­mentators) personally believe that they are seeking the truth about the world and interpret current work realistically, although it is not the best way to describe the aims of science, in my opinion. I do object to what I have termed “strong realism.”

ii. Several authors agree that global scientific realism is problematic, since there are fields such as quantum mechanics where realism is difficult to defend. These authors therefore retreat to local realism. But you go overboard in the opposite direction, ending up as a global nonrealist and therefore a global skeptic about science. Isn’t this a “retro” move that ignores the differences that local realists have noted?

Local realism is admittedly a difficult issue for me and (I suspect) for many others.

Although a nonrealist my position is very far from global skepticism. On the contrary, it is based on comparative reliability claims as well as forward-looking heuristic appraisals. To say that we have good empirical evidence that item A works better than item B for purpose P or that option O is currently more promising than option R does not express a skeptical position. I have no problem with pragmatic claims that we can genuinely justify, but I am indeed skeptical of strong realist claims that no future research will seriously affect present “mature” results.

My realist critic would like to say that I am making assumptions about the future in order to bias the case against realism. However, I am happy to leave the question of the future completely open. My point turns the tables on the objector, for it is the realists, not I, who are making far stronger claims about the future than they seem to realize. Treating all of future history as unknown and unknowable is quite under­standable, but to make strong claims about it is quite another matter.

b. Truth, reference, and pragmatic instrumentalism

i. Abandoning truth as a goal throws everything into chaos. We admire science as that human endeavor (alongside the law) that strives for objectivity and honesty and attempts to minimize error, to detect and correct it when it occurs. Denying that we can know scientific truth plays into the hands of the anti-science crowd, the purveyors of scientific uncertainty.

The objector confuses truth with truthfulness.^[60] While I am often agnostic about theoretical truth claims, I fully support truthfulness in the sense of honesty, clarity, and openness to criticism.

ii. You seem to agree with Richard Rorty (1991) that truth is only an empty compliment that we pay ourselves when our scientific claims satisfy the usual standards of consistency, high confirmation, and the like.

Yes, I believe that Rorty is basically correct on this point.^[61] In deep theoretical matters we have no direct access to the truth. Hence it cannot play a directly guiding role in research. When people are tempted to speak of truth in theoretical contexts it is because the methodological indicators that are available to us (predictive success, problem-solving effectiveness, fertility, etc., plus many lower-level checks in sci­entific practice) point in favorable directions. In my view (which may also have been Rorty's), truth-talk is a fagon de parler for reliability and for historically-indexed claims that X is the best theory or model or explanation (or whatever) yet available. Talk of truth is a sort of heuristic short cut!

My own pragmatic orientation again comes into play here. One reason why the Copernican dream of complete objectivity can never be realized (Nickles 2016) is that the dream is a moving target. Over time our scientific interests change, our goals and standards change. The natural world shapes our interests, to be sure, but it is we humans who decide where our research priorities lie. In that respect, we remain in control. A second reason is that frontier research is full of satisficing moves, as that great pragmatist, Herbert Simon, argued (e.g., Simon 1983).

iii. But surely you believe in well-established theoretical entities? For example, in the decades around 1900 there was a serious debate about whether atoms and molecules exist. French physicist Jean Perrin was able to show that they do exist and are responsible for the random jostling of visible particles in a liquid. Moreover, Perrin was able to establish the value of Avogadro’s number—the number of molecules in one mole of a substance—in many different ways. This robust result convinced even noted skeptics that atoms really exist. Surely you can’t deny that Perrin produced a strong case for the reality of atoms and molecules!

Perrin's work was very impressive. It provided the best warrant then available for taking seriously the postulation of atoms and molecules and the statistical apparatus developed by Boltzmann and others, including Einstein, for dealing with large pop­ulations of such particles. However, Perrin's work did not really establish what atoms are, and so the depth of the warranted realist conclusion was quite limited (cf. van Fraassen 2009). Under some conditions the “particles” posited by the best physicists are not even particles, and some such particles do not obey classical Maxwell-Boltzmann statistics. In these and many other cases of impressive theoretical and experimental work we lack basic understanding of “what is really going on”—a phrase that I find quite useful in the realism discussion. And, yes, over the rest of the century scientists learned a great deal about atoms, molecules, and chemical bonds, and so on. However, even today, there is much that we don't know about the exact nature of chemical bonds, even when we can manipulate the chemicals quite reliably.

I am not arguing that what we now claim to know is wrong. Rather, I am making the two-fold point, first, that such knowledge claims remain a far cry from “good old fashioned realism” (another useful phrase) that claimed that we were discov­ering—and understanding—what is really going on; and, second, that as science advances it gets stranger and stranger, i.e., harder than ever to understand. In this regard, scientific progress has not been in the direction of genuine realism. Core physics is where realists often take their stand on “mature” science. It is not a wise choice, for, insofar as they are correct that quantum theory (quantum field theory, the standard model, etc.) will last, they undermine their own strong realist position.

Do we really know what a molecule is? Molecules are made of atoms, which are made of electrons, protons, etc., which are made of quarks. Do we really know what an electron is in the sense of good old fashioned theoretical realism? The deeper and more “mature” our best science becomes, the weirder and less intelligible are the items it postulates in the old-fashioned sense of realism.

c. History, historicity, and progress

i. You say that you are a “complete historicist” who gives equal attention to future history and to past history. Letting pass that ‘future history’ sounds oxymoronic, there are several difficulties with your position. First, future history does not yet exist, so there is little we can say about it that bears on realism.

Admittedly, historians themselves are not in the business of forecasting the future and are professionally hesitant to talk about it. My point, again, is that strong realists themselves make much stronger claims about the future of science than some of them seem to realize. For if they truly believe that a deep theoretical claim is true, or so near the truth that only a bit of tinkering would make it true, then that theoretical domain has reached its terminus, the end of its dynamic history. In other words, strong realists predict (at least tacitly) that there will be (or need be) no significant changes in that domain in the entire future of scientific research. Sci­entists thousands of years in the future will still recognize these results as correct.

But such a claim is not really a scientific prediction, I would claim, but a forecast or even a prophesy. And, as noted above, forecasts of future knowledge and technical capabilities tend to be ludicrous, even fifty to a hundred years out. Yes, there is little that we can now say about the distant knowledge future, but the strong realist does say things that bear quite strongly on future history. Strong realists are acting as prophets, not warranted scientists.

ii. Second, you exacerbate the first difficulty by asking us to think about science a thousand years or even tens of thousands of years into the future. That is surely a silly exercise.

Why? History is (or will be) history, and realists are making claims about the far future, as we have just seen. It is their problem, not mine. I hereby confess to being an extreme historicist, a complete historicist.

iii. Third, appeals to possible future historical change do not raise specific objections to scientific claims today and therefore should not cast doubt on present science. Only actual difficulties need be taken into account.

The grain of truth here is that working scientists can address only actual (or concretely imaginable) difficulties. But strong realism is not working science in this sense. Strong realists cannot simply write off the future as irrelevant. Thus work such as Kyle Stanford’s on unconceived and underconceived alternatives (including those that will never happen to be conceived in the future) has considerable bite (Stanford 2006). The future remains highly contingent. It is rife with uncertainty, not merely risk (in the decision-theory sense). No one has carte blanche to ignore the future. Heightening sensitivity to future history—historicizing the future—is, indeed, a problem for all of us, for this very reason. We have so little information that is concrete enough to make the future “come alive” in the way that our past is (or was) alive; but we must do the best we can.

In Bayesian approaches to science the difficulty of establishing the truth or high probability of a hypothesis, given that only a few serious hypotheses have yet been formulated in a given domain, is known as the problem of the “catchall” hypothesis —the collection of all alternatives to the one being tested. Here is Wesley Salmon’s comment on the problem (Salmon 1991, 329):

What is the likelihood of any given piece of evidence with respect to the catchall? This question strikes me as utterly intractable: to answer it we would have to predict the future course of the history of science. No one is ever in a position to do that with any reliability.

iv. Fourth, despite your claimed neutrality about the future, your own position is, to some degree, based on your anticipations of what the future will bring. In particular, you seem committed to Thomas Kuhn’s claim in The Structure of Scientific Revolutions (1962) about occasional scientific revolutions in perpetuity.

That is a possibility that cannot be ignored, but I am not committed to it. It is also possible that the sciences will evolve slowly; and it is a basic lesson from evolu­tionary biology that very slow evolution can be as transformative as you please, given enough time. What we may call the evolutionary illusion undercuts both the revolutionary view and the strong realists’ static view. Evolutionary change can be so slow that it is hard to recognize within a single, productive, human lifetime. The evolutionary scale here is obviously much faster than much biological evolution, yet it is still very different from the scale of ordinary human experience. Note that the realist response to Kuhnian revolution claims—that there was, in fact, a good deal of continuity between the predecessor and successor theory or paradigm—does nothing to address the long-term evolutionary point. You can have all the continuity you want between temporally adjacent work, but over a long enough time the changes can be radical.

Strong realists (those who claim that we know that we already have the truth, or something very close) are committed to denying that the conditions for evolution (any longer) apply to mature fields. Yet all it takes for evolution to occur are variations, selections from among those, some of which are retained as a basis for future variation. Darwin himself pointed out that when these conditions occur (with the right sort of linkage), evolution is not improbable; on the contrary, it is virtually unstoppable.

v. How can nonrealists account for scientific progress? Hilary Putnam (1975, 73) once claimed that “Realism is the only philosophy that does not make the success of science a miracle. ” Even if the “miracle ” language is too strong, nonrealists and antirealists surely have a hard time explaining scientific suc­cess, whereas realists do not.

Here I must be extremely brief. I first note, on the negative side, that Putnam (1988) later abandoned this strong scientific realism for what he called “internal realism.” Second, I agree with those commentators who hold that the whole problem of explaining success is a pseudo-problem insofar as it seeks a monolithic answer in terms of Truth or a specially truth conducive Method unique to science. Scientists engage in all sorts of imaginative moves and checks on them. I repeat that it is strong realists who carry the burden of explanation if success means approx­imation to the truth, since nonrealists do not accept this strong view of progress. Third, I agree with Arthur Fine (1986), Bas van Fraassen (2002), and others that the miracle argument is fatally defective, or at best too weak to support the strong claims made by realists. As Fine noted early on, nonrealists do not regard apparent explanatory success as sufficient reason to conclude truth, owing to underdeter­mination, etc., yet the inference to truth as the best explanation for the success of science makes a similar move at the metalevel. Fourth, it is also gratuitous for those trying to understand how science works, since appeal to truth provides no new research tool.

Here an ambiguity in “explaining the success of science” emerges. Are we talking about the success of a particular theory or model? Then the realist faces the challenge of explaining the success of Newtonian theory, given how wrong we now think it is. Are we talking about the research activities of scientists, activities such as improved experimental design, better instrumentation, integration that makes possible more cross-checking and hence more robust results, more rigorous reviewing of publications, etc.? Appeal to truth is no help here. Or are we asking the transcendental question: What are the necessary presuppositions for a given science to have gotten progressively better over time? Strong realists appeal to truth in this latter sense, which (if I and other critics are right) has nothing to do with how science works in the previous sense. In saying that approach to the truth is the only, or most important, explanation of progress, realists are transcending working science.

To state the difference between the two positions in dramatic form, nonrealists want to know “the particular go” of science^[62] whereas strong realists want to believe that they have seen the Forms before they die. The trouble with this second desire, based on the idea that science has been moving toward the truth over a succession of long-term developments, is that the basic ontologies of these developments have changed radically (even among contemporary experts) and do not resemble suc­cessive approximation to a final, stable truth.

On the positive side, I suggest that we think of scientific progress in terms of ongoing research fertility and pragmatic progress in craft traditions and in trans­lational research rather than in terms of progress toward a universal, final truth. We have ways of recognizing improvement, innovation, when it comes to opening new research domains and design of new products and services and their modes of production and distribution, without needing to bring in truth with a capital T. Why can't we evaluate scientific products and processes in the same way? In point of fact, I suggest that we already do, often without realizing it. (Realists will reply that technical progress also calls for a truth explanation, and so we fail to engage again, across the two sides of the above ambiguity.) Pragmatic success must indeed engage the world,^[63] but success is also conditioned by human goals and standards as conditioned by global, local, and personal histories.

George Reisch (2016, 15) quotes from a letter Thomas Kuhn wrote to his mentor, James Bryant Conant, who urged Kuhn to drop the paradigm language from a draft of what would become The Structure of Scientific Revolutions: “Would you say that home industry was merely a less effective way of doing what the factory system later did?” Perhaps not, but my view remains closer to Kuhn's than to Conant's.

d. Alleged practical virtues of realism

i. Realists want to know the truth about reality. For them research into what really exists is an inspiring quest that energizes them. By contrast, nonre­alism dampens curiosity about our universe and thus diminishes motivation for research into the unknown. This was one of Popper’s objections to instrumentalism Popper (1963, Chap. 3).

First, I am not rejecting intentional realism, which is the form of realism that Popper had in mind. Popper was a nonrealist in my sense, given his denial that we can ever know the theoretical truth. If anything, he was a stronger agnostic than I am. Second, scientists and science analysts who are nonrealists are still attempting to find the best answers available to questions that they (we) find important; and we will continue to do so as curious people can conduct free inquiry. Progressives are always trying to improve on current capabilities. In fact, I believe that the position I defend encourages talented investigators to be bolder, to treat the future as still open to significant change, even in currently mature fields. It encourages people, from experts to the general public, to regard the sciences as an ongoing process of inquiry rather than as a body of established truth, thus in a more prospective and a less retrospective manner.

ii. Realism helps with what you call heuristic appraisal and is not opposed to it, as you seem to think. Crucial decisions at the research frontier involve which claims are stable enough to use as a basis for other work rather than being the focus of additional testing. Thus realism provides a good guide for what is hot and what is not. Moreover, finding what is real helps to constrain the other parts of the research. In short, realism enables us to distribute research resources more efficiently than does non-realism.^[64]

We get the same results with ordinary experimental and theoretical justification. Calling the result real does not add any epistemic warrant that was not already there. It is just a verbal label that we attach to mark an already made decision to change focus to something now considered more fruitful. Besides adding nothing to the research process, realism poses a danger of changing the focus too soon or too much, thereby discouraging fundamental research and thereby possibly distorting other work.

e. Postmodernism and policy concerns

i. You describe yourself as a pragmatist. Rorty dismisses the debate between realists and nonrealists (“constructivists”) on pragmatic grounds, as a difference that does not make a difference (Rorty 1991, 2; 2007, 34).

While I have some sympathy for Rorty's position (see a-ii above), I think he goes too far here.^[65] There is a difference between strong realism and nonrealism (and also antirealism) that can make a difference, both to science policy and to scientific practice in the research choices that scientists make. For example, funding agencies are unlikely to pour massive funding into areas declared to be certifiably true, thus at the end of their innovative history and hence sterile in terms of fundamental research (Nickles 2009, 2017). These also tend to be the most expensive specialties to support. And even accomplished scientists shift to new problem areas that they judge both more challenging and more fertile, once their old specialty seems to have reached its current goals.

ii. Realism provides a simplified model of science for public understanding of sci­ence and against the anti-science naysayers on evolution, climate change, and much else. The public is fascinated by scientific discovery because they believe it really is deep discovery of the wonders of our universe. To reject realism threatens to reduce public support for science, including government support.

There is a grain of truth to the point about public interest in science, but even among nonrealists there is great interest in learning the latest ideas about what our universe might be as well as in finding out what our universe is not. This last, a sort of “negative realism,” is something that I can accept. It is basically Popperian realism: denial that we are likely to know the positive theoretical truth in deep domains while being confident that some deep theoretical claims have been refuted. For example, useful as it is, I readily agree that our universe is not Newtonian, that it is shockingly weird compared to the classical universe.

iii. Opponents of scientific realism are often thought to be intellectually “soft, ” having fallen into radical (as opposed to moderate) postmodern social con­structivism and relativism.

No doubt some are. But it seems to me that the situation is largely just the reverse—that it is strong and deep realists who cannot fully resist the temptations of the cognitive-historical illusions and emotional psychological factors such as the satisfaction of predictive success and of neat problem solution, factors that go beyond the hard-nosed respect for empirical information and clear logical and mathematical thinking that the sciences, above all, are supposed to honor.

Aknowledgment I would like to thank Marlo Alai, Marco Buzzoni, Wenceslao Gonzalez, and Tomasz Placek for helpful conversation.